Inspired by Albert Einstein [1879-1955]
Learn how to identify anomaly within several similar objects with Artificial Intelligence
Working with time-series sensor generated data
Understand how Unsupervised Machine Learning Algorithm works using real life dataset
Learn developing in R and ShinyApp with a possibility to better explore the data, instantly deploy your project
Explained use of Version Control to be organized and save time
Practice with real life generalized Dataset coming from Manufacturing!
Versatile method is presented using a Case Study approach.
This method helped to discover real life inefficiency and to solve the problem!
Start with R here! Step by step introduction with examples and practice
Basic understanding on Time-Series data manipulation in R
More approaches of Anomaly Detection including Deep Learning on h2o framework is covered in the course
Practical Developing the idea of Industrial Process Control with Artificial Intelligence with DEMO Shiny Application included
Course video captions are translated to [Chinese-Simplified, Hindi, German, French, Italian, Portuguese, Turkish, Spanish, Malay, Indonesian, Russian] languages
Problem-solving in Manufacturing is usually perceived as a slow and boring activity especially when many possible factors involved. At the same time it’s often common that problems going on and on unobserved which is very costly. Is it possible to apply Artificial Intelligence to help human to identify the problem? Is it possible to dedicate this boring problem solving activity to computer? Apparently yes!!!
This course will help you to combine popular problem-solving technique called “is/is not” with Artificial Intelligence in order to quickly identify the problem.
We will use data coming from four similar Machines. We will process it through the Unsupervised Machine Learning Algorithm k-means. Once you get intuition understanding how this system work You will be amazed to see how easy and versatile the concept is. In our project you will see that helped by Artificial Intelligence Human eye will easily spot the problem.
Course will also exploit different other methods of Anomaly Detection. Probably the most interesting one is to use Deep Learning Autoencoders models built with help of H2O Platform in R.
Using collected data and Expert Knowledge for Process Control with AI:
In this course we will build and demo-try entire multi-variables process supervision system. Process Expert should select dataset coming from the ideally working process. Deep Learning model will be fit to that specific pattern. This model can be used to monitor the process as the new data is coming in. Anomaly in the process then can be easily detected by the process operators.
Ready for Production:
Another great value from the Course is the possibility to learn using ShinyApp. This tool will help you to instantly deploy your data project in no time!!! In fact all examples we will study will be ready to be deployed in real scenario!
You will learn R by practicing re-using provided material. More over you can easily retain and reuse the knowledge from the course – all lectures with code are available as downloadable html files. You will get useful knowledge on Version Control to be super organized and productive.
Join this course to know how to take advantage and use Artificial Intelligence in Problem Solving
To know what you will use in the course
In this lecture you will get introduced to the case study we will be looking at the course and the final result we will going to get
Quick intro on how to get the most of the course
The theory about Problem Solving and how can we combine that with Artificial Intelligence
To get intuitive understanding on what k-means clustering algorithm does
How to install R and R-Studio IDE
Learn to create your first projects in R. Also some little practice to code in ShinyApp
To know how to apply our method of Anomaly Detection directly in ShinyApp
In this lecture we will discuss about three main ideas for data loading to the ShinyApp
Lecture explaining manipulation of data using Pipe
Explained how to use inputs previously generated in ui.R inside server.R
Manipulation of data and performing Unsupervised Machine Learning with kmeans function in R
This lecture introduces industrial environment and the data coming from it
In this lecture we will see in the detail how possible app can look like. Potential solution explained will not be the final one. You can decide to apply different strategies to solve a business problem - alert operators of the process on up-coming anomaly!!!
This lecture is explaining one possible solution, covers possible use cases of this method in real industrial process. Setting up new challenges to achieve the same goal... feel free to contribute to Bonus Section of this course!!!
Bonus lecture about how to perform "data massaging" before applying Unsupervised Machine Learning. We will be working with Time-Series data in R
Deep Learning with H2O Machine Learning Platform. Part 1.
This lecture is dedicated to:
Lecture that focuses on
Introducing our working environment and new Einstein's quote!
Explaining the dataset for this Section
Setting up the demo shiny app. Train and save the model
How to use this shiny app for demonstration purposes
Additional article summarizing anomaly detection methods